Digital twin of a pipe conveying fluid: flow rate anomaly detection and quantification via multi-fidelity kalman filters and event based cameras signals

Numerical results of a digital twin (DT) framework for a pipe conveying fluid recorded using two event-based cameras (EVB). The DT framework consists of an unscented Kalman filter (UKF) for parameter identification and a linear Kalman filter (LKF) coupled with dynamic mode decomposition (DMD) for real-time monitoring and anomaly detection.

The test case concerns a 75-second recording with a time interval of 0.025 s for a change in flow velocity in the pipe carrying the fluid occurring at t = 18.4 s. The flowrate is the unknown parameter and is initialized as 5.84 m/s and is then reduced to 5.21 m/s. Details of the implementation are available in the associated article.

Uncontrolled Keywords

• Unscented Kalman Filter
• Dynamic Mode Decomposition
• Digital Twin
• Pipe conveying Fluid
• System identification
• Anomaly Detection
• Event-based reconstruction